whats interesting is the last comment All this happens over localhost, so it's quite fast, but in the | interest of efficiency, is there something I can set or call so that | it closes that first connection almost immediately? Having to refetch | a 800meg file off of NFS might hurt -- even if a good chunk of it is | still in the OS block cache.
You'd need to do this using inline C, but yes, anything is possible. (Sorry, I don't have an example for it here) What do you need to do via inline C to prevent the full 800 MB from being downloaded even the first time? On Mon, Mar 14, 2011 at 6:51 PM, Mark Moseley <[email protected]> wrote: > On Mon, Mar 14, 2011 at 3:30 PM, Chris Hecker <[email protected]> wrote: > > > > Anybody have any ideas? They're not all the same mime type, so I think > > putting them in an uncached dir is better if there's no way to figure it > out > > in vcl. > > > > Chris > > > > > > > > On 2011/03/13 07:26, AD wrote: > >> > >> i dont think you can check the body size (at least it seems that way > >> with the existing req.* objects ). If you know the mime-type of the > >> file you might just be able to pipe the mime type if that works for all > >> file sizes ? > >> > >> I wonder if there is a way to pass the req object into some inline C > >> that can access the body somehow? > >> > >> On Sat, Mar 12, 2011 at 11:28 PM, Chris Hecker <[email protected] > >> <mailto:[email protected]>> wrote: > >> > >> > >> I have a 400mb file that I just want apache to serve. What's the > >> best way to do this? I can put it in a directory and tell varnish > >> not to cache stuff that matches that dir, but I'd rather just make a > >> general rule that varnish should ignore >=20mb files or whatever. > >> > >> Thanks, > >> Chris > > > I was asking about the same thing in this thread: > > http://comments.gmane.org/gmane.comp.web.varnish.misc/4741 > > Check out Tollef's suggestion towards the end. That's what I've been > using. The one drawback is that it's still fetched by varnish > *completely* in the first, not-yet-restarted request, which means that > a) you're fetching it twice; and b) it'll still stored albeit > momentarily, so it'll evict stuff if there's not enough room. > > Before that, I wasn't sending any reqs for anything matching stuff > like .avi or .wmv to varnish (from an nginx frontend). > > It'd be kind of neat if you could do a call-out and for anything > matching a likely large file (i.e. has extension matching .avi, .wmv, > etc), and do a HEAD request to determine the response size (or > whatever you wanted to look for) before doing the GET. > > _______________________________________________ > varnish-misc mailing list > [email protected] > http://www.varnish-cache.org/lists/mailman/listinfo/varnish-misc >
_______________________________________________ varnish-misc mailing list [email protected] http://www.varnish-cache.org/lists/mailman/listinfo/varnish-misc
